Enantioselective Synthesis of the Cyclopiazonic Acid Family Using Sulfur Ylides

Abstract A convergent, nine‐step (LLS), enantioselective synthesis of α‐cyclopiazonic acid and related natural products is reported. The route features a) an enantioselective aziridination of an imine with a chiral sulfur ylide; b) a bioinspired (3+2)‐cycloaddition of the aziridine onto an alkene; and c) installation of the acetyltetramic acid by an unprecedented tandem carbonylative lactamization/N−O cleavage of a bromoisoxazole.

Prep-HPLC Where necessary, reverse-phase preparative HPLC was performed on Waters AutoPurification system with ACE 5 C18 columns (250×4.6 mm for method development, 250×21.2 mm for preparative runs), using acetonitrile-water gradients. NMR Routine NMR spectra were recorded on Varian, Bruker and JEOL spectrometers at 400 MHz for 1 H and 100 MHz for 13 C spectra. High-resolution spectra were run on Bruker Cryocarbon 500 spectrometer (500 MHz for 1 H and 125 MHz for 13 C). Signals are reported relative to the residual signal of the non-deuterated solvent (CDCl3: δ = 7.26 ppm, CD3CN: δ = 1.94 ppm for 1 H spectra; and CDCl3: δ = 77.16 ppm, CD3CN: δ = 118.26 ppm for 13 C spectra). 1 H NMR data are reported as follows: integration, chemical shift (parts per million, ppm), multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, dd = doublet of doublets, m = multiplet, br = broad, app = apparent), coupling constant (Hz) and description. IR Infrared spectra were recorded on a PerkinElmer Spectrum One ATR FT-IR: spectrometer as thin films. Only selected peaks are reported. Data represented as follows: frequency of absorption (cm −1 ), and intensity of absorption (s = strong, m = medium, w = weak, br = broad. HRMS High-resolution mass spectra (HRMS) were recorded by the University of Bristol Spectrometry Services Laboratory using electrospray ionization (ESI; Bruker micrOTOF II) and chemical ionization (CI; VG AutoSpec) techniques.

Melting Points
Melting points were determined using a Reichert hot stage apparatus with a digital thermometer. SFC Chiral supercritical fluid chromatography (SFC) was performed on a Waters TharSFC system using Whelk-O 1 column (4.6×250 mm, 5 µm), and monitored using a diode array detector (DAD). Crystallography X-ray diffraction experiments on compounds 25 and 33 were carried out at 100(2) K on a Bruker APEX II CCD diffractometer using Mo-Kα radiation (λ = 0.71073 Å). Intensities were integrated in SAINT [2] and absorption corrections were based on equivalent reflections using SADABS. [3] Structure 25 was solved using Superflip [4,5] while 33 was solved using ShelXT, [6] both of the structures were refined against F 2 in SHELXL [7,8] using Olex2. [9] All of the non-hydrogen atoms were refined anisotropically. All of the hydrogen atoms were located geometrically and refined using a riding model. In the case of 25 one of the NO2 group O and the acetonitrile solvent molecules in the lattice displayed disorder, the occupancies of the fragments was determined by refining them against a free variable with the sum of the two sites set to equal 1, the occupancies were then fixed at the refined values. Restraints and constraints were used to maintain sensible geometries and thermal parameters. In 33, Squeeze within Platon [10,11] was used to remove disordered solvent from the lattice that could not be sensibly modelled. Crystal structure and refinement data are given in Tables 8-9. Crystallographic data for compounds 25  Line broadening was applied at 0. 3 Hz, followed by the baseline correction (polynominal fit, 1 Hz filter). The reference peaks were integrated and the purity of the sample was estimated using the following equation: [12] ܲ where sample and ref refer to the sample and reference material parameters, respectively. I is the NMR integral of the signal. N is the number of protons in the signal. MW is the molecular weight. M is the mass. P is the purity.
Unless stated otherwise, all other commercially available reagents were used as received. 4-Bromo-3-formylindole was purchased from Frontier Scientific, Inc. and used as received. Pd(dppf)Cl2·DCM was purchased from Strem Chemicals, Inc. and stored at 23 °C in a desiccator. TsCl was recrystallized from hexane and stored at 23 °C under nitrogen. Hoveyda-Grubbs 2 nd generation catalyst (97%) and Grubbs 2 nd generation catalyst were purchased from Sigma-Aldrich and stored at 4 °C under nitrogen. Ti(OEt)4 (technical grade, 90%) was purchased from Sigma-Aldrich and used as received. In(OTf)3 was purchased from Sigma-Aldrich and stored at 23 °C in under nitrogen. N-bromosuccinimide (NBS) was recrystallized from water [16] and stored at −20 °C under nitrogen. Triflic acid was purchased from Sigma-Aldrich and used as a stock solution in anhydrous DCM (0.1 M), stored in a Schlenk tube under nitrogen.

General procedure
The starting aziridine cis-24 (15 mg, purified by HPLC) was placed into a flame-dried Schlenk tube, followed by naphthalene (internal standard, scintillation grade, 6−7 mg, weighed precisely). The tube was back filled with nitrogen. Anhydrous solvent (1 mL of DCM, unless indicated otherwise) was then added and the mixture was cooled to the starting temperature with dry ice bath (−70 °C, unless stated otherwise). The Lewis/Brønsted acid was then added (solids were added neat, liquids were pre-dissolved in anhydrous DCM to 0.2 M) and the reaction mixture was allowed to slowly warm up, while being stirred in the ice bath (warming rate: −70→−10 °C over 6 hr). Reaction aliquots were sampled as indicated and analyzed by LCMS: where IntSM is the current integral of the starting material measured vs. internal standard where IntTM is the current integral of the target material peak, IntIS is the current integral of the internal standard, Int 0 SM is the initial integral of the starting material, and Int 0 IS is the initial integral of the internal standard.   The starting aziridine 32 (purified by LCMS, 20 mg, 0.028 mmol) was placed into a flame-dried tube (8×90 mm), followed by a crystal of naphthaline (scintillation grade, internal standard). The tube was back filled with nitrogen. Anhydrous DCM (1 mL) was added and the tube was cooled to −78 °C. The promoter was then added (TfOH [0.1 M in DCM] or In(OTf)3 [neat], see Table 4) and the reaction progress monitored by LCMS (Agilent InfinityLab Poroshell 120 EC-C18, 2.7 um, 3.0×50 mm, 50→90% MeCN-water, 0.5 mL/min). For comparison of the reaction traces under TfOH and In(OTf)3 catalysis, see p. SI-8. Table 4. Optimization of the (3+2)-cycloaddition of cis-aziridine 32 Note. Trans-32 refers to aziridine mixtures enriched in trans-isomer (typically 3:1 for the racemic route and 9:1 for the enantioselective route).

Comparison of TfOH and In(OTf)3 in promoting the (3+2)-cycloaddition
The relative area under the curve (AUC) for the target material is essentially identical under the two conditions:  The N-Ts group was removed under the previously published conditions. [17] HPLC-grade THF, MeOH, and water were degassed (by sonication under vacuum for 5 min) individually. N-Ts α-CPA imine 28 (20 mg, 0.041 mmol) was placed into a 1.75 mL vial with a stirbar, followed by Cs2CO3 (53 mg, 0.16 mmol, 4 eq). Degassed solvents (0.50 mL of THF, 0.50 mL of MeOH, and 0.05 mL of water) were added, the overhead space was flushed with argon, the vial tightly capped and heated in a 65 °C oil bath, with daily monitoring of the reaction progress by LCMS (50-90 MeCN-H2O). After 4 days (approx. 90 hr), the conversion was deemed complete.
Yield: 9.5 mg (70%, d.r. 2.5:1)    Notes: * Commercial α-CPA from Alfa Aesar (J61594). Signals assigned based on 2D NMR data (COSY, HSQC, HMBC). # Lit. δC [18] Obs. δC ∆(δC) Lit. δH [18] Obs. δH ∆(δH)  # Lit. δC [18] Obs. δC ∆(δC) Lit. δH [18] Obs. δH ∆(δH)  A. These signals could not be observed by 13 C NMR directly due to broadening (C-5, 6,7,8,19) or being hidden under the solvent peak (C-22) and, where appropriate, were inferred from HSQC/HMBC. B. These signals were misassigned in the isolation report. [18] We reassigned them based on the HSQC and HMBC data of the synthetic material. The starting aldehyde 18 (500 mg, 1.36 mmol) and 4-nitrobenzenesulfonamide (330 mg, 1.63 mmol, 1.2 eq) were placed into a flame-dried Schlenk tube and backfilled with nitrogen. Anhydrous DCM (6.8 mL, 0.2 M) was then added. Neat Ti(OEt)4 (90%, 0.63 mL, 2.7 mmol, 2.0 eq) was added, and the reaction mixture was stirred at 23 °C until LCMS indicated complete consumption of the starting material (60−90 min). The reaction mixture was transferred into a 100 mL Erlenmeyer flask and diluted with 20 mL DCM. Water (30 mL) was added and the mixture was shaken for 5 sec, then gently swirled, to ensure the hydrolysis of [Ti] species. The resulting slurry was quickly filtered through a loose plug of glass wool (w×h: 2.5×5 cm), washing the solids with 20 mL of DCM and 20 mL of water. The resulting liquid mixture, mostly free from solids and slime, was transferred into a separating funnel. Note: if the filtration is not performed, the subsequent extraction becomes very inefficient and time consuming; also, if too tight a filter is used, it gets blocked almost instantly. The layers were separated, and the aqueous layer was extracted with DCM (3×10 mL). The combined organic layer was washed with brine (30 mL), dried over Na2SO4 and concentrated to give yellow solid, which contained the desired product along with 10−20% of NsNH2. The crude was transferred into a 50 mL round-bottom flask as a solution in 5 mL of DCM and stirred at 500 rpm. n-Pentane (5 mL) was added dropwise to precipitate impurities (mostly NsNH2), and after 10 min the solution was filtered through a plug of Celite (w×h: 1.5×2 cm), washing with 20 mL of 1:1 DCM-n-pentane. The volatiles were evaporated to give the product as yellow foam, which was used as is, without further purification (~90% purity by NMR). Small amounts of analytically pure samples could be prepared by recrystallization from 2:3 MTBE-heptane. Note: the product is extremely unstable of silica or alumina; all attempts to perform chromatography on it resulted in partial or complete decomposition. 1-((4-(Ethoxycarbonyl)-3-methylisoxazol-5-yl)methyl)tetrahydro-1H-thiophen-1-ium trifluoromethanesulfonate, 15a [19] N O

Comparison of NMR Data in Acetone
15a 22 23 The starting alcohol 21 (1.00 g, 5.4 mmol) and 2,6-di(t-butyl)-4-methylpyridine 23 (1.33 g, 6.49 mmol, 1.2 eq) were dissolved in anhydrous DCM (50 mL) in a flame-dried two-neck 250 mL flask and cooled to 0 °C. Then, Tf2O (1.09 mL, 6.49 mmol, 1.2 eq) was added and the clear colorless reaction mixture was stirred at 0 °C. In 5 min, white precipitate was observed. After 1 hr at 0 °C, the reaction was quickly passed through a 1.5×5 cm pad of silica into a flame-dried flask, washing the silica with 50 mL of anhydrous DCM. The filtrate was concentrated to give yellowish waxy solid of the intermediate triflate 22, which was used immediately in the next step. Triflate 22 was suspended in anhydrous Et2O and cooled to 0 °C. Tetrahydrothiophene (0.57 mL, 10.9 mmol, 2 eq) was added dropwise to the rapidly stirred solution and the resulting suspension was stirred at 0 °C for 1 hr. The product was then isolated by filtration, washed with 50 mL of Et2O and dried under high-vacuum.
White solid.
The starting alcohol 21 (100 mg, 0.54 mmol) and 2,6-di-tert-butyl-4-methylpyridine 23 (133 mg, 0.65 mmol, 1.2 eq) were placed into a flame-dried Schlenk tube, followed by anhydrous DCM (5 mL). The clear colorless solution was cooled to 0 °C. Triflic anhydride (110 µL, 0.65 mmol, 1.2 eq) was then added and the reaction was stirred at 0 °C for 1 hr. The resulting mixture was quickly filtered through a 1. Purification on Biotage Isolera system (ZIP KP-Sil 10 g, 1→20% MeOH-DCM, 1-10-2 CV) afforded the target material as white crystalline solid.  A 500 mL three-neck flask was charged with 4-bromo-3-formylindole (10.00 g, 44.64 mmol), then evacuated and refilled with nitrogen 3 times. Degassed (nitrogen sparging for 30 min) anhydrous THF (110 mL) was then added, followed by solution of KOH in degassed water (2.0 M, 33 mL, 67 mmol, 1.3 eq; prepared by dissolving 6.4 g of KOH in 50 mL degassed water (nitrogen sparging for 30 min) and cooling to 23 °C). Then, Pd(dppf)Cl2·DCM (1200 mg, 1.47 mmol, 3.3 mol%) was added in one portion. A reflux condenser was attached, the overhead space flushed with nitrogen (3 times), and the reaction mixture was placed into an oil bath (65 °C) and stirred for 20 min. A solution of allyl-Bpin (10.9 mL, 58 mmol, 1.3 eq) in anhydrous THF (20 mL) was then added with a syringe pump (9.9 mL/hr). After all allyl-Bpin was added, the reaction mixture was stirred for another 20 min, when TLC analysis showed complete consumption of the starting material. The reaction mixture was cooled to 23 °C, and diluted with 200 mL of ether and 100 mL of water. The phases were separated, and the aqueous layer was extracted with ether (3x30 mL). The combined organic layer was washed with brine (50 mL), dried over Na2SO4, and concentrated to give crude indole as deep-red oil. The crude material was immediately dissolved in anhydrous DMF (120 mL) in a flame-dried 500 mL flask and cooled to 0 °C. Then, NaH (2.50 g, 60% in oil, 62.5 mmol, 1.4 eq) was added portionwise (warning: intense gas evolution!) When bubbling ceased (~10 min), TsCl (11.0 g, 58 mmol, 1.3 eq) was added in one portion. The reaction mixture was stirred at 0 °C for 20 min and at 23 °C for 4 hr. The reaction mixture was then cooled to 0 °C and the reactants carefully quenched with 100 mL of water. The mixture was then poured into a mixture of 50 mL water, 100 mL of brine and 100 mL of ether. The layers were separated and the aqueous layer was extracted with ether (6x60 mL). The combined organic layer was washed with brine (100 mL), dried (Na2SO4) and concentrated to give dark brown oil. Purification on Biotage Isolera system (ZIP KP-Sil, 120 g, 5 → 35% EtOAc-petrol, 1-10-2 CV) provided the target material as brownish solid.  The reaction mixture was diluted with 10 mL of DCM, the volatiles were evaporated, and the crude mixture was purified on Biotage Isolera system (45 g KP-Sil + 5 g Thelos NM loading cartridge, 5 → 35% EtOAc-hexane, 3-10-2 CV) to give the product as off-white amorphous solid. Note: the starting material 17 is poorly soluble in 2-methyl-2-butene, and thus must be ground prior to use, to ensure complete conversion.   The known isoxazole 21 was prepared using a modified literature procedure. [20,21] Furanone 42 (5.50 g, 32.3 mmol, 1.0 eq) was dissolved in anhydrous EtOH (30 mL), followed by NH2OH·HCl (2.25 g, 32.3 mmol, 1.0 eq) and NaOAc (2.65 g, 32.3 mmol, 1.0 eq). The resulting suspension was heated at reflux for 80 min, when TLC (60% EtOAc-petrol) indicated complete consumption of the starting material. Volatiles were evaporated. The residue was diluted with half-sat. NH4Cl (100 mL) and extracted with 3×30 mL of Et2O. The combined organic layers were washed with brine (30 mL), dried over Na2SO4 and concentrated. Drying under high-vacuum to remove traces of EtOH gave pale brown waxy solid, which was used as received (note: contains ~2% of isoxazole regioisomer). An analytically pure sample was prepared by chromatography on Biotage Isolera system (100 mg of crude 21 on 10 g SNAP Ultra cartridge, 10 → 50% EtOAc-petrol, 1-10-2 CV).

15a
Ns Sulfonium salt 15a (526 mg, 1.3 mmol, 1.3 eq) and anhydrous Cs2CO3 (423 mg, 1.3 mmol, 1.3 eq) were placed into a flame-dried Schlenk tube. Anhydrous DCM (14 mL) was added and the yellowish suspension was stirred at 0 °C for 1 hr, then at 23 °C for 1 hr, then recooled to −40 °C. Neat imine 14 (550 mg, 0.99 mmol, 1.0 eq) was added in one portion, and the tube walls were washed with 2 mL of anhydrous DCM. The reaction mixture was stirred at −40 °C for 2 hr, then at 0 °C for 2 hr. The reactants were quenched at 0 °C with 5 mL of half-sat. NaCl. The reaction mixture was then subjected to extraction (30 mL water -3×20 mL DCM). The combined organic layer was dried over Na2SO4, and concentrated to give brown-orange foam. The crude material was analyzed by quantitative NMR and used as is.
Trans:cis = 9:1. The trans-isomer is highly unstable and isomerizes into the cis-isomer under acidic conditions (e.g. SiO2, CDCl3). The trans/cis ratios should be measured within 10 min of NMR sample preparation. Solutions in C6D6 are reasonably stable.
Note: Analysis of NMR spectra is greatly complicated by the existence of the E-and Z-isomers as inseparable 1:1−1.5:1 mixtures. The 1 H NMR signals could be assigned to the individual isomers only with moderate confidence. The 13 C NMR spectrum is reported as is.   The starting isoxazole 35 (3.00 g, 26.7 mmol) was dissolved in reagent grade acetic acid (130 mL) in a 250 mL round-bottom flask, followed by N-bromosuccinimide (NBS, 5.67 g, 31.9 mmol, 1.2 eq) and conc. H2SO4 (2.8 mL, 53 mmol, 2.0 eq). The brownish reaction mixture was stirred at 23 °C for 5 min, until most of solid dissolved. Then, a reflux condenser was attached, the overhead space was exchanged with nitrogen (3 times), and the reaction was Step 1. The starting alcohol 29 (100 mg, 0.52 mmol) and 2,6-di(t-butyl)-4-methylpyridine 23 (128 mg, 0.63 mmol, 1.2 eq) were dissolved in anhydrous DCM (5 mL) in a flame-dried Schlenk tube and cooled to 0 °C. Then, Tf2O (0.10 mL, 0.63 mmol, 1.2 eq) was added and the resulting clear colorless solution was stirred at 0 °C. In 5 min, white precipitate was observed. After 1 hr at 0 °C, the reaction mixture was quickly filtered through a 1.5×1 cm (i.d./h) pad of silica (pre-washed with 12 mL of anhydrous DCM) into a flame-dried flask (25 mL, pear-shaped), washing the filter cake with 7 mL of anhydrous DCM. The filtrate was concentrated to give yellowish-white waxy solid of the intermediate triflate 30, which was used immediately in the next step.
Step 2. The triflate 30 was suspended in 4 mL of anhydrous Et2O in the same flask, cooled to 0 °C (note: the compound is only partially soluble in this solvent), and rapidly stirred at 600 rpm. Neat tetrahydrothiophene was added dropwise, and the resulting white suspension was stirred at 0 °C for 1 hr. The solids were then filtered and washed with 20 mL of ether to provide the crude product as off-white wax. Purification by column chromatography (2 → 20% MeOH-DCM) gave the target material as white powder.  Step 1. The starting alcohol 29 (250 mg, 1.30 mmol) and 2,6-di(t-butyl)-4-methylpyridine 23 (320 mg, 1.56 mmol, 1.2 eq) were dissolved in anhydrous DCM (9 mL) in a flame-dried Schlenk tube and cooled to 0 °C. Then, Tf2O (0.26 mL, 1.56 mmol, 1.2 eq) was added and the resulting clear colorless solution was stirred at 0 °C. In 5 min, white precipitate was observed. After 1 hr at 0 °C, the cold reaction mixture was passed through a 1.5×1 cm (i.d. × h) pad of silica (pre-washed with 12 mL of anhydrous DCM) into a flame-dried flask (50 mL, pear-shaped), washing the filter cake with 30 mL of anhydrous DCM. The filtrate was concentrated to give off-white waxy solid of the intermediate triflate 30, which was used immediately in the next step.
Step 2. The triflate 30 was suspended in 10 mL of anhydrous Et2O in the same flask, cooled to 0 °C (note: the compound is only partially soluble in this solvent), and rapidly stirred at 600 rpm. A solution of (+)-39 (143 mg, 0.57 mmol, 1.1 eq) in anhydrous Et2O (3 mL) was added dropwise, and the resulting suspension was stirred at 0 °C for 2 hr. After 2 hr at 0 °C, the precipitate was filtered off, washed with 30 mL of Et2O and dried in air. Purification on Biotage Isolera system (SNAP Ultra 10 g, 2 → 20% MeOH-DCM, 1-10-2 CV, loaded in DCM) provided the desired product as off-white powder. Notes: 1. The filter cake in step 1 needs to be washed with sufficient amount of DCM to ensure complete elution of triflate 30. 2. Failure to pre-dissolve the sulfide (+)-39, or too rapid addition of it, leads to coagulation of the reaction mixture, poor yields and hard-to-purify mixtures. The starting alcohol 29 (300 mg, 1.56 mmol) and 2,6-di(t-butyl)-4-methylpyridine 23 (384 mg, 1.88 mmol, 1.2 eq) were dissolved in anhydrous DCM (12 mL) in a flame-dried Schlenk tube and cooled to 0 °C. Then, Tf2O (0.32 mL, 1.9 mmol, 1.2 eq) was added and the resulting clear colorless solution was stirred at 0 °C. In 5 min, white precipitate was observed. After 90 min at 0 °C, the solids were removed by filtration through a 1×3 cm (i.d./h) pad of silica (pre-washed with 22 mL of anhydrous DCM) into a flame-dried 50 mL flask, washing the filter cake with 20 mL of anhydrous DCM. The filtrate was concentrated to give yellowish-white waxy solid of the intermediate triflate 30, which was used immediately in the next step. The triflate 30 was suspended in anhydrous Et2O (12 mL) and cooled to 0 °C. A solution of the chiral auxiliary 40 (0.32 mL, 1.9 mmol, 1.2 eq) in anhydrous Et2O (4 mL) was added dropwise and the resulting mixture was stirred at 0 °C for 4.5 hr. The white suspension was then carefully filtered, attention being given not to transfer any of the brown wax covering the flask walls. The filter cake was washed with Et2O (20 mL) and dried under vacuum to give the product as ivory powder. Notes: 1. It is important to add the auxiliary 40 slowly as a solution in Et2O, otherwise the reaction mixture is prone to aggregation and the product purity suffers. 2. During the reaction, the product crushes out as nice free-flowing white solid, while the unreacted triflate 30 and the side-products remain stuck to the flask walls as brown wax.
3. The compound 31c is unstable in CDCl3, acidic, or protic solvents (t½[CDCl3] ~90 min). The decomposition product is alkene 41 (cf. ref [ [23,24] ]). Our early attempts to prepare 31c in the DCM-water mixtures led to the exclusive formation of 41. 4. We could not identify the means to purify 31c, therefore it is paramount that it is generated in sufficient purity straight from the reaction. To a solution of 38 (2.35 g, 9.22 mmol) in 18 mL of degassed DCM were added tetrahydrothiophene (2.44 mL, 27.7 mmol), degassed water (6 mL) and NaBF4 (3.04 g, 27.7 mmol), sequentially. The resulting biphasic mixture was stirred vigorously (1000 rpm) at 23 °C under nitrogen for 48 hr. It was then diluted with water (10 mL) and DCM (50 mL). The two layers were partitioned, and the aqueous layer was extracted further with DCM (3×15 mL). The combined organic layer was dried over Na2SO4, filtered and concentrated to give a white solid. The crude product was dissolved in 30 mL of DCM and added dropwise to rapidly stirring Et2O (250 mL). The flask containing the crude product was rinsed with additional DCM (2×5 mL) and added to Et2O. An immediate white solid formation was observed. The mixture was stirred in an ice-water bath for 30 min and then kept chilled and unstirred for 2 h. The white solid was filtered (gravity), washed with Et2O and dried in vacuo to afford sulfonium salt 31d (2.276 g, 70%) as a white solid.
The starting imine 14 (410 mg, 76% pure by qNMR, 0.57 mmol) and chiral sulfonium salt 31b (335 mg, 0.68 mmol, 1.2 eq) were placed into a flame-dried Schlenk tube and back-filled with nitrogen. Anhydrous MeCN (8 mL) was added and the brown solution was cooled to −20 °C (Cryostat). Flame-dried K2CO3 (156 mg, 1.13 mmol, 2.0 eq) was then added and the reaction mixture was stirred at −20 °C for 22 hr. The reactants were quenched with water (30 mL) and the mixture was then extracted with DCM (4×20 mL). The combined organic layer was washed with brine (20 mL), dried over Na2SO4 and concentrated to give brown foam. The product was used crude without further purification. A 10-mg sample was used for qNMR analysis, and another 10-mg sample subjected to prep-HPLC and subsequent chiral SFC analysis.

Purification
The trans-aziridine trans-32 is highly unstable and converts into cis-isomer cis-32 in CDCl3, MeCN/water, or on SiO2/DCM. Analytically pure cis-aziridine was routinely prepared by ageing the crude trans/cis-mixtures in CDCl3 (6-22 hr), and then purifying by preparative reverse-phase HPLC (80−95% MeCN-water over 25 min; tR: cis 11.09 min, trans 11.83 min). Attempted separation of the trans-32 by prep-HPLC led to the isolation of ~1:1 trans/cis-mixtures, presumably as the result of isomerization in the presence of a protic solvent (water).
For characterization data, see below.
Illustration of the structure of (±)-33 with atomic numbering scheme depicted. Only one of the two unique molecules in the asymmetric unit is shown for clarity, the relative orientations of all groups are the same in both molecules. Thermal ellipsoids depicted at the 50% probability level and hydrogen atoms omitted for clarity. A solution of acetaldehyde oxime (mixture of syn and anti-isomers, 1.00 g, 16.9 mmol) and propargyl alcohol (2.0 mL, 34 mmol, 2.0 eq) in 60 mL of DCM was cooled to 0 °C and stirred for 10 min. Aqueous NaOCl (14.8 wt%, 14 mL, 34 mmol, 2.0 eq) was then added dropwise over 20 min. The reaction was allowed to warm up to 23 °C gradually and stirred overnight (17 hr) under nitrogen. It was then diluted with 10 mL of water, and the two phases were partitioned. The aqueous layer was extracted with DCM (2×20 mL), and the combined organic layer was dried over MgSO4, filtered and concentrated. Purification by flash column chromatography (33→50% EtOAc-hexane) afforded alcohol 35 as a volatile light-yellow oil.
Light-yellow oil.
TLC (33% EtOAc-n-pentane): Rf = 0.13 (UV inactive, stains with KMnO4). A solution of alcohol 35 (3.36 g, 29.7 mmol) in 60 mL of anhydrous DCM was cooled to ca. −17 °C in a MeOH-ice bath, and stirred for 10 min under nitrogen. PPh3 (9.35 g, 35.6 mmol, 1.20 eq) and CBr4 (11.8 g, 35.6 mmol, 1.20 eq) were then added sequentially. The resulting dark orange mixture was stirred at this temperature for 1 hr, and then at 23 °C for 3 hr, at which time TLC analysis indicated full consumption of alcohol 35. All volatiles were removed in vacuo. Purification by flash column chromatography (14 → 20% EtOAc-n-pentane) afforded 37 (4.48 g, 86%) as a pale yellow oil. The starting isoxazole 37 (1.00 g, 5.68 mmol) was dissolved in reagent grade acetic acid (50 mL), followed by Nbromosuccinimide (NBS, 1.21 g, 6.8 mmol) and sulfuric acid (95%, 0.61 mL, 11 mmol), in a 100 mL recovery flask, to give a clear colorless solution. A reflux condenser was attached, the setup was briefly back filled with nitrogen (5 times) and the reaction mixture was heated at 110 °C for 1 hr. After cooling, the mixture was diluted with 10 mL of Et2O and stripped of volatiles at 30-33 °C. The residue was diluted with DCM (10 mL) and remaining reactants were quenched with 40 mL of sat. Na2CO3 (caution: intense gas evolution). The layers were separated and the aqueous layer was extracted with 3×10 mL of DCM. The combined organic layer was washed with brine (20 mL), dried over Na2SO4 and concentrated to give cloudy pale-brown oil. The crude mixture was purified by silica chromatography on a Biotage Isolera system (SNAP KP-Sil 25 g, 2 → 30% EtOAc-petrol, 1-10-2 CV) to give a colorless oil, which solidified (white, waxy solid) upon standing in freezer (ca. −20 °C).
TLC (10% Et2O-n-pentane): Rf = 0.37.  The known furanone 42 was prepared following a literature procedure. [21] Magnesium ethoxide (9.36 g, 82 mmol, 1.1 eq) was placed into a flame-dried 100 mL flask, followed by anhydrous ethanol (1.7 mL), anhydrous benzene (18 mL), and reagent-grade ethyl acetoacetate 20 (10.0 g, 74.7 mmol, 1.0 eq). The mixture was stirred at 23 °C for 90 min, then cooled to 0 °C. Anhydrous MeCN (18 mL) was added, followed by chloroacetyl chloride (6.5 mL, 82 mmol, 1.1 eq) with vigorous stirring (~600 rpm). The reaction mixture was manually swirled for 30 sec, then allowed to warm up to 23 °C over 3 hr with magnetic stirring. The mixture was poured into 50 mL of ice water containing 3 mL of conc. H2SO4. The layers were separated and the aqueous layer was extracted with 3×20 mL Et2O. The organic layer was washed with brine (20 mL), dried over MgSO4 and filtered. The resulting solution was cooled to 0 °C. TEA (10.3 mL) was added to the rapidly stirred solution, to precipitate acidic side-products. The stirring was continued at 0 °C for 30 min. The solids were then filtered off, washing with Et2O. The clear solution was concentrated to give brown oil.
The product was purified on Biotage Isolera system (3 portions, ZIP KP-Sil 120 g, loaded neat, 20 → 100% Et2Opetrol), collecting the second major peak. The resulting yellow solid was used in the next step. A small, analytically pure, sample was prepared by crystallization from hot Et2O.
White film.